#38100·openvino

[错误]: 在 NPU DEVICE_ARCHITECTURE=4000 (Lunar Lake) 上, LFM2.5 350M 和 1.2B 会在 NPU 上默默地产生垃圾 — 在 3720、CPU 和 GPU 上正确

作者: aweussom创建于 2026年9月13日更新于 2026年9月18日
标签PSEcategory: NPU

import sys import openvino_genai as ovg, openvino as ov dev = sys.argv[1] if len(sys.argv) > 1 else "NPU" core = ov.Core() if dev == "NPU": print("arch", core.get_property("NPU", "DEVICE_ARCHITECTURE"), "| driver", core.get_property("NPU", "NPU_DRIVER_VERSION"))

huggingface-cli download OpenVINO/LFM2.5-350M-int8-ov --local-dir ./lfm2.5-350m

kwargs = {"MAX_PROMPT_LEN": 1024} if dev == "NPU" else {} # NPU-only property pipe = ovg.LLMPipeline("./lfm2.5-350m", dev, **kwargs) cfg = ovg.GenerationConfig(); cfg.max_new_tokens = 40; cfg.do_sample = False for p in ("Say hello.", "What is 2+2?", "What is the capital of Norway?"): pipe.generate(cfg, p) print(pipe.generate(cfg, "What is the capital of Norway?"))

内容来源: openvinotoolkit/openvino